In the course of the past five seasons my early practices were revolutionised. My antipathy to snow was overcome, and my predilection for rocks was modified. Like all those who are not mountaineers born, I was, at the first, extremely nervous upon steep snow. The snow seemed bound to slip, and all those who were upon it to go along with it. Snow of a certain quality is undoubtedly liable to slip when it is at a certain inclination.(153) The exact states which are dangerous, or safe, it is not possible to describe in writing. That is only learnt by experience, and confidence upon snow is not really felt until one has gained experience. Confidence gradually came to me, and as it came so did my partiality for rocks diminish. For it was evident, to use a common expression, that it paid better to travel upon snow than upon rocks. This applies to snow-beds pure and simple, or to snow which is lying over glacier; and in the selection of routes it has, latterly, always been my practice to look for the places where snow slopes, or snow-covered glaciers, reach highest into mountains.(154) It is comparatively seldom, however, that an ascent of a great mountain can be executed exclusively upon snow and glacier. Ridges peep through which have to be surmounted. In my earlier scramblings I usually took to, or was taken upon, the summits (or arêtes) of the ridges, and a good many mountaineers habitually take to them on principle, as the natural and proper way. According to my experience, it is seldom well to do so when any other course is open. As I have already said, and presently shall repeat more particularly, the crests of all the main ridges of the great peaks of the Alps are shattered and cleft by frost; and it not unfrequently happens that a notch in a ridge, which appears perfectly insignificant from a distance, is found to be an insuperable barrier to farther progress; and a great detour, or a long descent, has to be made to avoid the obstacle. When committed to an arête one is tied, almost always, to a particular course, from which it is difficult to deviate. Much loss of time must result if any serious obstruction occurs; and defeat often follows a temporary check. But it rarely happens that a great alpine peak is seen that is cut off abruptly, in all directions, from the snows and glaciers which surround it. In its gullies snow will cling, although its faces may be too steep for the formation of permanent snow-beds. The merits of these snow-gullies (or _couloirs_) have been already pointed out,(155) and it is hardly necessary to observe, after that which was just now said about snow, that ascents of snow-gullies (with proper precautions) are very much to be preferred to ascents of rocky arêtes.